Someone should teach the decision makers how pipelines work. If AI-created diffs are being churned out at 10x the previous rate but manual reviews are the bottleneck then the overall system is producing at the exact same rate as before. The only thing you have added is cost, uncertainty and engineers being less familiar with the system.

Check back in a year or two. Just the other day we had Claude finding significant bugs in Firefox. https://news.ycombinator.com/item?id=47273854

The next thing these geniuses will think of will be to have AI review the diffs.